[Snowman]

I've actually done a lot of research on this topic because I'm a nutjob who collects data on anything and everything.

While it is true that you can find cards in old slabs that would regrade the same today, those cards are certainly the exception, not the rule. To answer the question posed in the OP, you really need to look at the statistical distributions of grades rather than looking at one-off cards. Also, many of the long-time collectors with impressive vintage collections have large quantities of cards with old cert numbers. They often defend them out of bias (even if it's a subconscious bias). Typically, when you see an older cert that would grade the same today, it is because of the centering requirements. Like with Phil's beautiful 64 Topps Hank Aaron posted above. Many of these cards could have graded higher at the time if they had better centering, but the centering held them back. Those cards would grade the same today if the corners looked better than the assigned grades but centering held them back. The centering standards haven't really shifted, at least not as PSA. Not by much (although the variance in results from one grader to the next has gotten worse).

Here is what the data I have collected shows. The data was gathered by me looking at thousands of cards in holders and grading them myself (mostly post-war vintage, although I think it's fair to assume these observations apply to pre-war as well), by PSA's & SGC's published grading standards. Then I compare my grade against the grade on the slab and align them to when they were graded.

Early PSA cert numbers (00xxxxxx to ~15xxxxxx & some of the certs that were used at shows and various events like 3000xxxx, 4000xxxx, 5000xxxx, and 9000xxxx) all have a median grading delta distribution of about 1 grade more generous than I would give them (that's the center of the distribution - think of a bell curve - yay Central Limit Theorem!).

Less early, or somewhat more recent, cert numbers (~15xxxxxx to ~2999xxxx) have a median grading delta distribution of approximately zero, which is to say they align with my grades pretty closely - the median & average grade differences that is. So, most cards that I would grade as a 5 or 6 are in fact graded as a 5 or 6 from this era. Hence the delta is centered around 0.

For more recent certs (~4100xxxx to 8xxxxxxx) the grades are much harsher. The center of the grading delta distribution is approximately one full grade tougher than I would assign. There's also more variance among these cert numbers with respect to the distributions, meaning the tails on the bell curve of the delta between my grades and the assigned grades are wider than it is for the other earlier cert numbers. This reflects the fact that PSA has gotten less consistent as they've had to hire graders rapidly to keep up with the demand in the pandemic era.

One thing you'll notice is that if you look at the high-grade vintage cards (9s and 10s) in pretty much every major auction, almost all of them have old cert numbers. This is telling. Also, the majority of them have, in my opinion, likely been altered/trimmed.

SGC has also gotten more strict over time, but not as much as PSA. Early SGC cards are about 1 grade more generous than my grades while current SGC grades and grades from about 8 years ago or so have delta distributions that are centered right around where my grades are. So they are in alignment with how I grade, currently. They also have shorter tails, so less variance (likely due to having a smaller grading team and just being better/more experienced graders overall than PSA - at least on average).

BGS/BVG - I have no idea, and I don't care. I don't track them.